Have you tried fitting the data (or the two different sets of data) to the MM equation using non-linear regression? Is it possible to reach Vmax for the low [S] range in your experiments (based on the linear fits it doesn't seem possible) ?

Your linear fits do illustrate quite nicely the effect you're outlining. I can't think of anything else than an allosteric effect, which is caused by the substrate. You could also see something like this if your substrate manifests as several species in solution from which some undergo catalysis easier (say, if the substrate is an acid with several conjugate bases with certain equilibrium constant(s) and pH in your experiment changes or your substrate is a chelating agent, which forms complexes with metals etc. and is thus incapable of undergoing catalysis at low [S] ).

As far as i remember there is no possible way in which a substrate act as an activator allosterically which would generate a bifasic behavior.

Maybe your ligand is rescuing a partially folded population of enzyme and that is the second saturation you see.

Arne and Felipe-

Thanks for you insight into our odd kinetics. We are definitely trying non-linear regression to get a better fit. You're right that it is not possible to experimentally derive a Vmax for the low [S] end of the range, but we can derive a "pseudo" Vmax by just using the four low [S] points. Another possibility has presented itself that has to be investigated by NMR or LC/MS. Our substrate is a very large/complex natural product, and we will need to establish if the two kinetic components represent the expected chemistry (low Km perhaps) and a 2nd "rogue" activity at a different location on the molecule. Felipes comments are appropriate, since these are 6His tagged proteins. Many years ago, I had previous experience with a different enzyme in which the His tag altered the kinetic properties of the enzyme significantly compared to untagged protein. I was unable to get good kinetics (appeared non-Michaelis) with the His-tagged enzyme, so opted to purify the untagged enzyme in order to get good kinetics. Thanks again for your insight.

Paul -

Interesting problem. I concur with your comment above and I wonder if you have a mixed population of substrates rather than a mixed population of enzyme states. In the past, I have encoutered difficulties substrates that contained small amounts of contaminants that had potent binding but poor turnover that resulted in a similar readout. Your case is pretty extreme in that they are so clearly resolved into two distinct phases. You noted in your comment above that the substrate is a complex mixture - do you see batch-dependent kinetic profiles using the same enzyme prep and conditions? This would be a clear flag to chase down the substrate rather than the enzyme.

Paul Swanson, if you are sure that you are dealing with a single enzyme, than I think that you are right about the possibility of an allosteric site in the enzyme, and this is modulating the activity either through the substrate or the product. Whether the modulator is the substrate or the product, will be indicated by the kinetic data. The question does not carry any information about the substrate and product as well as any co-factors and other reaction condition.